Wavelet-based FLD for face recognition

We introduce, in this paper, a face recognition method which employs the wavelet-based FLD (Fisher Linear Discriminant). This paper deals with accuracy as well as facial speed for face recognition. An image of 128/spl times/128 is decomposed to sub-images with 16/spl times/16 resolution by wavelet transform. Then two sub-images of low and mid-range frequency bands are trained and used to recognize faces. Experimental results show that the proposed algorithm is faster maintaining the recognition rate of original FLD method. Our experiments, the computational time is reduced by factor of about 6.

[1]  Lawrence Sirovich,et al.  Application of the Karhunen-Loeve Procedure for the Characterization of Human Faces , 1990, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  M. Turk,et al.  Eigenfaces for Recognition , 1991, Journal of Cognitive Neuroscience.

[3]  Abdesselam Bouzerdoum,et al.  Illumination invariant face recognition , 1998, SMC'98 Conference Proceedings. 1998 IEEE International Conference on Systems, Man, and Cybernetics (Cat. No.98CH36218).

[4]  Shimon Ullman,et al.  Face Recognition: The Problem of Compensating for Changes in Illumination Direction , 1994, IEEE Trans. Pattern Anal. Mach. Intell..

[5]  David J. Kriegman,et al.  Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Linear Projection , 1996, ECCV.

[6]  Mayer Aladjem.  Linear Discriminant Analysis for Two Classes via Removal of Classification Structure , 1997, IEEE Trans. Pattern Anal. Mach. Intell..

[7]  Juyang Weng,et al.  Using Discriminant Eigenfeatures for Image Retrieval , 1996, IEEE Trans. Pattern Anal. Mach. Intell..

[8]  Keinosuke Fukunaga,et al.  Introduction to Statistical Pattern Recognition , 1972 .